Background

International efforts to control malaria supported by the Global Fund and the President's Malaria Initiative (PMI) are underway in many parts of Africa using strategies based on the scaling-up of long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) [1–3]. Pyrethroids are the only group of insecticides currently recommended for use on mosquito nets [4]. Pyrethroid resistance has, in recent years, become widespread among anopheline mosquitoes in western and southern Africa and also occurs in eastern and central Africa [5–8]. The recent evolution and spread of pyrethroid resistance in West Africa in the Mopti (M) molecular form of Anopheles gambiae sensu stricto presents a grave threat to control because carriers of this particular resistance are not killed by pyrethroid-treated nets or residual spraying [9, 10]. With the scaling-up of malaria control efforts the continued selection of pyrethroid resistance [11, 12] may compromise malaria control programmes and render this group of insecticides ineffective. It is important to investigate alternative insecticides on nets to supplement existing pyrethroid-based LLINs.

The target site insensitivity gene that confers knock down resistance kdr to pyrethroids in An. gambiae shows cross resistance to DDT [6]. Of the insecticides recommended by WHO for IRS, the most long-lasting and cost-effective is DDT [13]. No assessment of DDT has been made in areas where kdr is prevalent. A small-scale trial of DDT is essential before any decision to redeploy DDT can be made in West Africa.

Because of DDT's damaging environmental impact, the Stockholm Convention on Persistent Organic Pollutants stipulates that 'countries are encouraged to reduce and eliminate the use of DDT over time and switch to alternative insecticides' [14]. The alternative classes of insecticide to DDT - the organophosphates, carbamates and pyrethroids - are more expensive and shorter-lived. For programmes that use DDT to remain viable it is necessary to develop long-lasting formulations of the alternative insecticides before DDT can be replaced [15]. Dow Agrosciences has among its portfolio of insecticides the organophosphate chlorpyrifos methyl which is effective against anophelines but applied as a wettable powder, the standard IRS formulation, it is too short-lived [16]. The company has, therefore, developed a microencapsulated formulation to improve residual activity. Its limited environmental persistence [16] and lack of cross resistance makes chlorpyrifos methyl a more attractive prospect than DDT for IRS should re-formulation increase its residual life. It also has the advantage of not selecting for pyrethroid resistance or undermining LLIN use.

This paper reports on an experimental hut trial in southern Benin of microencapsulated chlorpyrifos methyl. Efficacy is compared with that of DDT and the pyrethroid lambdacyalothrin in an area where An. gambiae has become difficult to control with pyrethroids [9].

Methods

Study sites and experimental huts

The evaluation was carried out in experimental huts situated in Ladji, a peri-urban village on the periphery of Cotonou, the capital of Benin. The village floods during the rainy season, creating breeding sites for An. gambiae. The local An. gambiae population is comprised entirely of the M taxon and is resistant to pyrethroids and DDT, with kdr at high frequency and metabolic resistance also present [9, 17]. The nuisance mosquito Culex quinquefasciatus is present year round and is resistant to pyrethroid, carbamate and organophosphate insecticides [17]. The experimental huts are made from concrete bricks, with roofs of corrugated iron, ceilings lined with plastic sheeting on the interior surface and walls plastered with a cement/sand mix [18–20]. Each hut stands on a concrete base surrounded by a water-filled moat to exclude scavenging ants. Entry of mosquitoes occurs via four slits, 1 cm wide, located on three sides of the hut. Mosquitoes exit into a verandah trap projecting from the fourth side.

Insecticide treatments

The insecticides used were:

  • chlorpyrifos methyl 24% CS ('Reldan GF 1246', Dow AgroSciences)

  • DDT 50% WP (sourced by Dow from South Africa)

  • lambdacyhalothrin 2.5% CS ('Icon', Syngenta, Switzerland) microencapsulation designed for ITNs.

  • Lambdacyhalothrin 10% WP ('Icon' Syngenta, Switzerland) wettable powder designed for IRS.

The following treatments and application rates were compared in seven experimental huts:

  • chlorpyrifos methyl IRS at 500 mg/m2

  • DDT IRS at 2 g/m2

  • Lambdacyhalothrin 10% WP, IRS at 30 mg/m2

  • chlorpyrifos methyl ITN at 100 mg/m2

  • lambdacyhalothrin 2.5% CS, ITN at 18 mg/m2

  • unsprayed control hut

  • untreated net

The Chlorpyrifos methyl dosage applied for IRS was recommended by Dow Agrosciences whereas the application rate for the ITN was the one used in experimental huts in Ivory Coast to control the Savanah (S) taxon of pyrethroid-resistant An. gambiae[21].

The dosages for DDT and lambdacyalothrin treatments are those conventionally used and recommended by WHO.

The test nets were made of 100-denier polyester netting in which a total of 80 holes of 4 cm2 area were cut along each side and end panel to simulate wear and tear. Nets were treated with insecticide by hand dipping. Insecticide was sprayed onto interior walls and plastic sheeting using a Hudson compression sprayer equipped with a flat fan nozzle. The evaluation started one week after treatment and ran for 42 nights from 8 April to 24 June 2005.

Study procedure

The two net treatments and the untreated control net were rotated between three of the huts at six-day intervals. The four huts dedicated for IRS treatment were fixed throughout the study and the treatments could not, of course, be rotated. The volunteer sleepers gave informed consent and were provided with chemoprophylaxis. They slept in the huts from 20:00 to 05:00 each night, and were rotated between huts on successive nights to adjust for any variation in attractiveness to mosquitoes. Mosquitoes were collected each morning at 5:00 from floors, walls, ceilings and verandahs, and transported to the laboratory for identification to species, mortality counts and determination of gonotrophic condition. Live mosquitoes were held in plastic cups and delayed mortality was recorded after 24 h. The effects of each treatment were expressed relative to the control in terms of:

  • Deterrence: percentage reduction in the number of mosquitoes caught in treated hut relative to the number caught in the control hut;

  • Induced exiting: percentage of the mosquitoes collected from the verandah trap of treated hut relative to percentage caught in verandah trap of control hut;

  • Inhibition of blood-feeding: percentage of the mosquitoes collected which was blood fed in the treated hut relative to percentage blood-fed in the control hut;

  • Induced mortality: percentage of dead mosquitoes in treated hut relative to percentage dead in control hut.

  • If a treatment deters a significant number of mosquitoes from entering the hut, the values given by proportion blood feeding in the treatment hut may underestimate the full personal protective effect of the treatments. The personal protective effect of a treatment is better described by the reduction in the actual number of blood-fed mosquitoes in the treatment hut relative to the number blood-fed in the control hut:

where Bu = is the total number of blood-fed mosquitoes in the untreated control huts and Bt is the total number blood-fed in the huts with insecticide treatment.

Residual activity of insecticide treatments

To assess residual activity on treated walls or nets cone bioassay tests were undertaken each month using 3-5 day old An. gambiae females of a laboratory susceptible strain (Kisumu). Mosquitoes were exposed to nets for 3 min or to sprayed walls for 30 min as per WHO guidelines [22]. Approximately 50 mosquitoes were used per test.

Data analysis

Data were entered in Excel and transferred to STATA 6.0 software for further analysis. The numbers of mosquitoes collected each night and the actual numbers bloodfed was analysed per treatment using the Kruskal-Wallis and Wilcoxon rank sum tests. Proportional data (exiting rate, blood-feeding, mortality) were analysed using logistic regression after adjusting for the effect of sleeper and hut. A Chi-square test for trend was performed to assess whether there was any change in mortality over time in bioassay tests.

Results

Trials of ITNs against An. gambiae and Cx. quinquefasciatus

Table 1 shows the total number of An. gambiae and Cx. quinquefasciatus collected from the rooms with untreated or insecticide treated nets and the proportions exiting into the verandas. Figure 1A shows the percentage blood-fed and dying among the total collected. An average of 14 An. gambiae and 19 Cx. quinquefasciatus females were collected each morning from the rooms and verandas of the huts with nets. There was no significant difference in the number of An. gambiae collected between the untreated control and chlorpyrifos methyl huts (p = 0.49). By contrast, the number of An. gambiae found in the lambdacyhalothrin huts was 43.6% lower than in the untreated control huts (P < 0.001). Hence, there was no evidence of deterrence with the organophosphate chlorpyrifos methyl in contrast to the pyrethroid lambdacyhalothrin ITN. The trends with Cx. quinquefasciatus were similar to An. gambiae.

Table 1 Summary results of mosquitoes hut frequenting habit and exit rates in huts.
Full size table
Figure 1
figure 1

Percentage of Anopheles gambiae and Culex quinquefasciatus bloodfed and dead in experimental huts. Bloodfeeding (A) and mortality (B) rates with 95% confidence interval, of An. gambiae and Cx. quinquefasciatus in huts with different indoor residual spray (IRS) treatments and insecticide treated net (ITN). For each mosquito species within each intervention arm (IRS and ITN), treatments sharing the same letters in the middle of bars are not significantly different. LC = Lambdacyhalothrin; CM = Chlorpyrifos methyl; DDT = Dichlorodiphenyltrichloroethane

Full size image

The chlorpyrifos-methyl- and lambdacyhalothrin-treated nets induced little or no additional exiting of An. gambiae or Cx. quinquefasciatus into the verandas over that recorded for untreated nets.

The proportions of An. gambiae and Cx. quinquefasciatus blood-feeding through the sides or holes of the chlorpyrifos-methyl ITN were not significantly different from the untreated control nets or lambdacyhalothrin treated nets (P > 0.05 for both species). However, the reduced entry rate of An. gambiae to hut with lambdacyhalothrin-treated net indicated a 44.6% lower risk of exposure to potential infective bites than in huts with control or chlorpyrifos methyl-treated nets (table 1). Hence, there was no evidence of blood-feeding inhibition with chlorpyrifos-methyl- or lambdacyhalothrin-treated nets though the pyretroid on net still procures some appreciable degree of personal protection against pyrethroid-resistant An. gambiae or Cx. quinquefasciatus in this area of Benin.

The percentage mortality among An. gambiae was 45.2% with the chlorpyrifos methyl-treated net and only 29.8% with the lambacyalothrin-treated net (Figure 1B). Mortality rates among Cx. quinquefasciatus were lower than among An. gambiae and did not exceed 15% with either type of treated net (Figure 1B).

Trials of IRS against An. gambiae and Cx. quinquefasciatus

Table 1 shows the total number collected from the IRS treated rooms and the proportions exiting into the verandas. Figure 1B shows the percentage blood-fed and dying among the An. gambiae and Cx. quinquefasciatus collections. Owing to differences in site attractiveness of individual huts, which were located at different places within the village, and to the inability to rotate IRS treatments, it was not possible to interpret the overall numbers collected from the rooms in terms of treatment effects. Differences in hut position were a confounding source of error. By serendipity a significantly larger number of An. gambiae were collected from the chlorpyrifos-methyl treated hut than from the other types of hut (Table 1).

The percentages of An. gambiae and Cx. quinquefasciatus collected from the verandas of DDT, lambdacyhalothrin and chlorpyrifos methyl IRS treated huts were similar to those from the control verandas; the only exception was the significantly higher proportion of An. gambiae, which exited the lambdacyhalothrin treated hut (Table 1).

Blood-feeding rates of An. gambiae and Cx. quinquefasciatus in the chlorpyrifos methyl and DDT treated huts were not significantly different from the untreated control (Figure 1A). In the lambdacyhalothrin treated hut half as many Cx. quinquefasciatus were blood-fed compared to the control hut (P < 0.001). However a similar trend was not observed with An. gambiae in the lambdacyhalothrin treated hut.

Mortality of pyrethroid resistant An. gambiae was 95.5% with chlorpyrifos methyl IRS compared to 50.4% in the hut sprayed with DDT and 30.8% in the hut sprayed with lambdacyhalothrin (Figure 1B). The mortality of Cx. quinquefasciatus in the chlorpyrifos methyl IRS hut was 66.1% whereas in the DDT and lambdacyhalothrin IRS huts was only 14% (Figure 1B). Chlorpyrifos methyl IRS showed greater potential than DDT or lambdacyhalothrin IRS for control of pyrethroid resistant An. gambiae M form and Cx. quinquefasciatus in areas of high kdr frequency.

Neither chlorpyrifos methyl nor lambdacyhalothrin stained the sprayed surfaces nor did they cause an unpleasant odour or led to any complaints of adverse effects among the operators or sleepers at any stage of the study.

Residual activity

Figure 2 shows the mortality of An. gambiae freely entering the ITN and IRS treated huts divided into fortnightly intervals during the 42 days of the trial. The proportion of An gambiae killed in huts with chlorpyrifos methyl ITN during weeks 0-2 was 73% after which mortality showed a progressive decline over the remaining six weeks (P < 0.001) (Figure 2A). By contrast, chlorpyrifos methyl IRS consistently killed more than 95% of pyrethroid resistant An. gambiae throughout the trial period. Mortality rates of An. gambiae in huts with DDT IRS and lambdacyhalothrin ITN or IRS decreased steadily between weeks 2-8 (P < 0.0001), presumably due to loss of activity and survival of An. gambiae carriers of kdr or other forms of pyrethroid resistance.

Figure 2
figure 2

Mortality of Anopheles gambiae freely entering the ITN and IRS treated huts on a fortnightly basis. Anopheles gambiae mortality rates with 95% confidence interval, over fortnightly intervals during the hut trials with (A) insecticide treated nets (ITN) and (B) different indoor residual spray (IRS) treatments. Percentages are based on all mosquitoes collected from the rooms and veranda traps of huts. LC = Lambdacyhalothrin; CM = Chlorpyrifos methyl; DDT = Dichlorodiphenyltrichloroethane

Full size image

Figure 3 gives results of bioassay tests conducted with susceptible An. gambiae Kisumu strain on (A) ITNs and (B) IRS cement wall surfaces. The residual activity of chorpyrifos methyl on ITN confirmed the rapid decline in effectiveness, with the mortality rate decreasing from 100% to 9.7% within one just month (P < 0.0001) (Figure 3A). Lambdacyhalothrin ITN remained highly active (100% mortality) for up to 6 months before showing a decrease in mortality at month 9 (P = 0.008).

Figure 3
figure 3

Bioassay tests monitoring the residual activity of treatments using the susceptible Anopheles gambiae Kisumu strain. Monitoring of the residual efficacy under WHO cone tests of (A) insecticide treated nets (ITN) and (B) different indoor residual spray (IRS) treatments against susceptible An. gambiae Kisumu strain in experimental huts at Ladji. LC = Lambdacyhalothrin; CM = Chlorpyrifos methyl; DDT = Dichlorodiphenyltrichloroethane.

Full size image

The bioassays on chlorpyrifos methyl treated cement walls showed no loss of activity during the nine months of follow-up (P = 0.79) (Figure 3B). By contrast, decay of DDT and lambdacyhalothrin was evident on walls within the first month of spraying (P < 0.001).

Discussion

Of the various groups of insecticide recommended by WHO for indoor residual spraying [23] the most cost-effective, DDT, is compromised by its negative environmental impact, and the most widely used, the pyrethroids, will surely accelerate the selection of resistance and undermine that other great tool for malaria prevention, the long-lasting pyrethroid-treated net [9, 10, 24]. The ideal compound for IRS would come from an entirely different class of insecticide to pyrethroids or organochlorines. The formulations of organophosphates and carbamates currently recommended for IRS are relatively short-lived [23] and this had limited their deployment by malaria control programmes. In the present study, the microencapsulated chlorpyrifos methyl applied as IRS treatment in Southern Benin killed almost all pyrethroid and DDT resistant An. gambiae that entered the huts and demonstrated a residual activity that lasted for at least nine months without showing any decay detectable by bioassay. This constitutes an insecticide formulation to rival DDT and pyrethroids in terms of residuality and cost effectiveness.

Chlorpyrifos methyl CS has potential for deployment in a variety of epidemiological settings. In areas of West Africa where pyrethroids or DDT no longer control An. gambiae because of resistance [9, 10] chlorpyrifos methyl CS makes a promising alternative. Indeed, use of chlorpyrifos methyl CS or other long lasting cholinesterase inhibitor will be essential to keep malaria control or malaria elimination targets on track. With national malaria burdens continuing to fall [24] epidemics may become more frequent and IRS, as an emergency response may become more widely practised. Chlorpyrifos methyl CS is a less risky choice than DDT or pyrethroid if the local resistance status of An gambiae is unknown, because organophosphate resistance in this species is comparatively rare compared to pyrethroid resistance [17].

With its remarkable efficacy and residual activity it is essential to consider the issue of organophosphate resistance management from the outset. The most pragmatic approach to manage insecticide resistance is to rotate insecticides with differing modes of action between spray campaigns, although in reality sequential substitution of one unrelated compound for another once the former has failed is more the norm [25]. Ideally the use of pyrethroids should be constrained in order to preserve ITNs and LLINs. A promising partner insecticide to rotate with chlorpyrifos methyl in spray programmes is chlorfenapyr as this novel insecticide shows no cross resistance to OPs or pyrethroids in An. gambiae or Cx. quinquefasciatus, and has shown potential for IRS in areas where resistance to these two classes of insecticide has not yet been reported [26].

The short-lived residual activity of DDT in this trial (<2 months) compared to conventional wisdom (>6 months) might be due to the formulation obtained, to the cement substrate on which this insecticide was applied or to its residual activity perhaps being over stated [23]. Longer residual activity of DDT has been observed elsewhere on wooden walls in villages in Brasil [27] or Madagascar [28], but wood substrates are known to be rather benign to all types of insecticide. The study conducted in Southern Benin provides no reassurance that the activity shown by DDT would provide effective or sustainable control of malaria in areas where pyrethroid resistance involving kdr has become the norm.

There is a deterrent effect of lambdacyhalothrin-treated net against pyrethroid-resistant An. gambiae in Southern Benin and trustworthy higher personal protection than chlorpyrifos methyl because treatment could be rotated. For the lambdacyhalothrin IRS the result is less trustworthy because IRS treatments could not be rotated and the huts showed site effects with respect to mosquito numbers.

The hut trial also demonstrates the potential of chlorpyrifos methyl on nets for control of pyrethroid-resistant vectors. During the first few weeks, before chlorpyrifos methyl started to decay or wear off, the level of mortality was 100% against An. gambiae. With the advent of LLIN treatment technology there is scope for improving residual efficacy on nets using appropriate binding or incorporation technology.

Conclusion

With its good safety profile, low mammalian toxicity and residual activity, chlorpyrifos methyl meets the profile of a cost-effective replacement for DDT or pyrethroids. The challenge of finding an environmentally acceptable alternative to DDT appears to be met. Adopting a IRS strategy that incorporates chlorpyrifos methyl will reduce the selective pressure generated by pyrethroids and help preserve the future of LLINs.

Despite the great promise shown by chlorpyrifos methyl it seems unlikely that the manufacturer, Dow AgroSciences, will deploy the long-lasting formulation in the near future, because of uncertainties in the vector control market. The comparatively small size of the market and the unpredictability of winning tenders is a major deterrent to companies not already engaged in malaria control. The manufacturer should be encouraged by international donors and technical authorities to pursue further development and evaluation against malaria indicators in endemic settings where An. gambiae is pyrethroid-resistant or IRS is being considered for malaria control.

Authors' information

Raphael N'Guessan, Martin Akogbeto and Mark Rowland are all associated with the Pan African Malaria Vector Research Consortium, http://www.pamverc.org.